Research on the Prediction of Sunspot Numbers based on Neural Networks

Sunspots are strong magnetic fields on the sun, and the solar surface where sunspots appear is the active region of the sun. The number of sunspots reflects the strength of solar activity and is a key indicator of changes in solar activity. Predicting the number of sunspots is of great value for human space exploration and application activities. Meanwhile, the prediction of sunspot numbers has also played an important role in satellite navigation, communication and other fields. In the ionospheric IRI model, the number of sunspots is one of the input parameters. Therefore, studying the prediction methods of sunspot numbers, improving the accuracy and precision of predictions, is of great significance for ensuring the safety and normal operation of human activities. In this paper, a short-term prediction method for sunspot numbers based on neural networks is studied. Using historical observation data of sunspot numbers as training data, parameters are adjusted to obtain the optimal prediction network. The trained network is used to predict the number of sunspots in the next three days. Experimental results show that the method used in this paper can effectively predict the number of sunspots. When forecasting one day in advance, the mean absolute error of the predicted number of sunspots does not exceed 9, and when forecasting three days in advance, the mean absolute error does not exceed 16.